OVERALL - PROJECT SUMMARY The objective of the GENCODE consortium is to create a foundational reference genome annotation, in which all gene features in the human and mouse genomes are identified and classified with high accuracy based on biological evidence, and then to release these annotations for the benefit of biomedical research and genome interpretation. GENCODE aims for a better understanding of a `normal' human genome; using genome sequences of the most commonly used mouse strains will facilitate the most effective use of these key models for large-scale knockout analysis and disease-specific research. To produce regular annotation releases of high accuracy, GENCODE will continue to follow its well-established and conservative research design, supplemented by targeted investigations into the value of new technologies, new data and new sources of evidence. GENCODE focuses on protein-coding and non-coding loci, including their alternatively spliced isoforms and pseudogenes. Over the course of this proposal GENCODE will follow major directions in genomics, including graph- based genome representations, long-read transcriptome sequencing, connecting genes and the associated regulatory regions that affect their transcription, and identifying genes that are not present on the current reference assembly. The GENCODE consortium has four fundamental components: (1) a comprehensive gene annotation pipeline leveraging manual annotation; (2) an integrated approach to pseudogene identification and classification; (3) a set of computational methods to evaluate and enhance gene annotation; and (4) complementary experimental pipelines for validation and functional annotation. More specifically, in the next four years GENCODE aims to (1) extend the human and mouse GENCODE gene sets to as near completion as possible given current experimental technology; (2) deploy population-based genome annotation to ensure that any transcript isoform expressed in an individual human will be present in the reference annotation set; (3) extend the gene annotation to include core regulatory regions and tissue-specific enhancers from selected datasets; (4) to distribute GENCODE annotations and engage with community annotation efforts. Current popular distribution channels for GENCODE data including the GENCODE web site, the Ensembl and UCSC Genome Browsers, will be maintained. Finally, new mechanisms for prioritizing genes for manual annotation with community input will be established, with the long-term aim of establishing GENCODE as the standard annotation set for research and clinical genomics efforts.